What DeFi Total Value Locked Actually Measures (and What It Does Not)

Total value locked became the headline metric for decentralised finance. It is also one of the most frequently misread numbers in crypto. Here is what TVL captures, where it misleads, and why protocol quality matters more than size.

Where TVL comes from

Total value locked is the aggregate dollar value of all assets currently deposited in a DeFi protocol’s smart contracts. For a lending protocol such as Aave, it includes both the assets that depositors have supplied to earn interest and the collateral that borrowers have locked up to receive loans. For an automated market maker, it is the total value of all assets sitting in every liquidity pool. For a staking contract, it is the value of all tokens staked to earn rewards or secure the network.

DeFi Llama, the analytics site that has become the reference source for TVL data, aggregates these figures by querying the on-chain balances of smart contracts across more than fifty blockchains. Their methodology is publicly documented and their adapters are open source, which makes it possible to audit any protocol’s calculation independently. The markets page on TheWeal uses DeFi Llama figures for total DeFi TVL display.

What TVL is actually measuring

TVL measures the dollar value of assets inside smart contracts at a given moment. That is a reasonable proxy for the level of trust and usage a protocol has attracted, and for the liquidity available to users. A lending protocol with $5 billion in TVL has $5 billion worth of assets that borrowers can draw on and that depositors are earning yield on. A DEX with $2 billion in liquidity can handle larger trades with lower slippage than one with $50 million.

However, TVL is denominated in dollars, which means it fluctuates with crypto prices even when nothing changes in the protocol itself. During the bull run that ended in late 2021, total DeFi TVL across all chains reached roughly $180 billion, according to DeFi Llama historical data. By the end of 2022, with prices sharply lower but protocol usage broadly unchanged, it had fallen to around $35 billion. The protocols had not lost 80% of their users; the assets they held had lost 80% of their dollar value.

This price-sensitivity means TVL charts look like crypto price charts. Investors citing TVL growth as evidence of fundamental improvement need to check whether the growth came from more deposits or from the price of existing deposits rising. The native-unit TVL — the amount locked measured in ETH, or in each token directly — strips out the price effect and gives a cleaner picture of whether adoption is actually growing.

The double-counting problem

DeFi protocols compose: the output of one protocol becomes the input of another. A user can deposit ETH into a liquid-staking protocol such as Lido to receive stETH, deposit that stETH as collateral into Aave to borrow USDC, and deploy that USDC into a yield protocol. The ETH appears in Lido’s TVL, the stETH appears in Aave’s TVL, and the USDC appears in the yield protocol’s TVL. The same underlying ETH has been counted three times.

DeFi Llama introduced a “double count” toggle that attempts to exclude this recursion, but it relies on protocols correctly flagging their tokens as derivative representations of underlying assets. Coverage is incomplete, and the distinction between double-counting and legitimate protocol composability is sometimes genuinely ambiguous. A protocol that issues a receipt token for deposited liquidity is not trivially double-counting — the receipt token has its own market and its own risk profile separate from the underlying.

For analytical purposes, the safest approach is to look at TVL by protocol in isolation, compare protocols within the same category (lending vs lending, DEX vs DEX), and use TVL as one data point alongside revenue, active users, and fee generation.

TVL and actual risk

A large TVL does not mean a protocol is safe. It means a large amount of money is trusting its code. If that code contains a bug, more money is at risk. Several protocols that ranked in the top ten by TVL at their peak were later exploited for hundreds of millions of dollars. The Wormhole bridge, which held over $1 billion in TVL, was exploited for $320 million in February 2022 via a signature verification bug. The Ronin bridge, which settled transactions for the Axie Infinity game, lost $625 million to a validator key compromise in March 2022.

Conversely, a protocol with modest TVL may have an excellent security track record and a model that works at scale but has not yet attracted the marketing attention that drives deposits. TVL measures size, not quality.

Which chains host the most TVL

Ethereum remains the dominant chain for DeFi TVL by a wide margin. Despite higher fees than competing chains, it has the longest track record, the most battle-tested code, the deepest stablecoin liquidity and the widest integrations with institutional custodians. Protocols with billions in TVL tend to deploy first or exclusively on Ethereum, even if they later expand to other chains.

BNB Chain, Solana, Arbitrum, and Base each host significant DeFi ecosystems with lower fees and faster confirmation times, though each involves different trust assumptions (BNB Chain’s validator set is small; Base is operated by Coinbase). Tron hosts a large stablecoin TVL driven primarily by USDT, though most of it sits in a single lending protocol rather than a diversified ecosystem.

For coins tracked on TheWeal’s DeFi category, we note the primary chains of operation because cross-chain TVL figures can be presented deceptively when a project counts liquidity on a chain where it has inflated its own deposits to manufacture apparent size.

Revenue as a better signal

Protocol revenue — the fees paid by users and retained by the protocol or its token holders — is a harder metric to inflate than TVL. Inflating TVL requires only depositing your own assets and withdrawing them before end-of-day accounting. Inflating revenue requires paying fees, which costs real money. DeFi Llama publishes protocol revenue alongside TVL, and comparing the two gives a rough capital efficiency ratio: how much revenue is generated per dollar of TVL.

Decentralised exchanges tend to generate higher revenue per TVL than lending protocols, because swap fees are earned on volume while lending protocols earn on outstanding loans (a fraction of TVL). A lending protocol with high TVL and low revenue may be attracting deposits by subsidising them with token emissions rather than from organic demand. That subsidy eventually runs out. The methodology section describes how TheWeal’s model accounts for protocol fundamentals in constructing market scenarios.

What to watch instead of just the headline number

TVL in native units (ETH or token-denominated) removes price noise. Revenue per unit of TVL signals capital efficiency. The ratio of borrowed capital to supplied capital in a lending protocol (the utilisation rate) shows whether liquidity is being put to work or sitting idle. Unique addresses depositing is a rough user count, though it is easily gamed by a single actor using many wallets.

None of these metrics is a perfect signal, and none replaces reading the protocol’s code or audit reports. But used together they paint a more accurate picture than TVL in dollars alone — which is what most DeFi headlines report.

Frequently asked questions

Sources

• DeFi Llama — Total Value Locked across all chains (DeFi Llama, accessed June 2026)
• DeFi exploit leaderboard — Rekt News
• Decentralised finance on Ethereum — ethereum.org

Analysis reflects publicly available data as of publication. This is not financial advice. DeFi protocols carry smart-contract, liquidity and regulatory risks. Past TVL levels are not indicative of future performance. Model-based scenarios. Not financial advice.